Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
1
Alfred Holl
The inflectional morphology of the Swedish verb with respect to reverse order:
analogy, pattern verbs and their key forms
Summary
Learners of a language often try to construct analogy rules based on similarity. In the case
of verbs, similarity can be defined with respect to reverse order of present infinitives.
However, similar Swedish verbs need not belong to the same conjugation class. With regard
to verbs ending in -a, there is a triple choice: e. g., skida belongs to the 1st, smida to the 2nd
and rida to the 4th conjugation. On the other hand, there are homogeneous groups whose
verbs all belong to the same conjugation, e. g., all verbs ending in -änna belong to the 2nd
conjugation. This phenomenon is a difficult hurdle for learners of Swedish. There is not any
grammar book which comprehensively shows them where they can trust in this kind of
analogy base. In this paper, this problem is dealt with from a descriptive, synchronic and
graphematic point of view. Verb tables are presented which comprise the inflectional
morphology of the entire Swedish verbal system. They meet four requirements: every
arbitrary Swedish verb (without any exception) can be assigned to its pattern verb by a
simple, right-bound, longest matching algorithm. Thus, it is shown where analogy rules are
applicable. The key forms of each pattern verb are recorded so that every other verb form can
be derived. The verb tables can easily be adapted to language changes.
1 Introduction
One core problem for the learner of the Swedish language is to find out which conjugation
class a given verb follows. This is quite difficult, as Swedish has four classes: 1st, 2nd (with a
couple of important exceptions in several subclasses) and 3rd are weak, 4th is strong (a result
of the Indoeuropean ablaut). In comparison, English and German only have two conjugation
classes each: a weak one and a strong one. Swedish 3rd conjugation verbs are (except for a
handful of 4th conjugation verbs) the only ones ending in vowels other than a. The other ones
all end in a. As you can find exhaustive lists of 4th conjugation verbs and of 2nd conjugation
exceptions in Swedish grammar books, but none of the regular 2nd conjugation verbs, the
biggest partial problem for the learner is to distinguish between 1st and 2nd conjugation. In
this paper, however, I deal with the entire problem.
The learner of Swedish can solve this problem by using a dictionary where the conjugation
class is indicated for every verb. How can he remember the conjugation classes for 10,000
Swedish verbs, however? He will try to find heuristic rules. Normally, these are based on
similarity, type construction and assumption of analogy, e. g.: sända has a similar
graphematic shape as tända and follows the same conjugation as tända (the 2nd conjugation).
Therefore, the beginner will intuitively assume the following rule: “All verbs of the type
«ending in -ända» should follow the conjugation of tända (the 2nd conjugation)”, i. e. “tända
should be the pattern verb for all verbs ending in -ända” and “this verb group should be
homogeneous with respect to conjugation”. This assumption is not correct, however, as ända,
blända, fullända, skända belong to the 1st conjugation. Parallels can be found in other
Germanic languages: e. g., English to like - liked - liked vs. to strike - struck - struck and
German siegen - siegte - gesiegt vs. liegen - lag - gelegen.
The mentioned lack of analogy is a frequent phenomenon in the Swedish verbal system.
Although their infinitives look quite similar, Swedish verbs can belong to different
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
2
conjugation classes. Depending on their assignment, verbs can even have different meanings:
e. g., sluta means ‘to finish’ in the 1st conjugation and ‘to close’ in the 4th conjugation. On the
other hand, there are some cases where analogy rules can be applied successfully, e. g., “all
compounds of tända follow the conjugation of tända” and “all verbs ending in -änna follow
the conjugation of känna”. This situation is not a problem for adult Swedish native speakers,
but it is one for Swedish children before school age as well as for foreigners. Up until now,
however, only few analogy rules and exceptions have been mentioned in Swedish grammar
books; a complete overview has not existed. Although teaching Swedish as a second
language became more and more important during the last 20 years, the estimation in
Hellberg 1978: 17 is still valid: “A great deal has been written about the inflectional
morphology of Swedish, but very few attempts at a fully comprehensive description have
been made.”
I would like to thank Torbjörn Fogelberg, a Swedish native speaker, who graduated in
Scandinavian philology from Lund University. He checked the entire paper and gave
important contributions. I, myself, know Swedish quite well, but I am not a native speaker.
2 Aims
With my approach, I address language teachers, language learners and linguists who are
interested in descriptive and normative grammar, particularly in inflectional morphology. For
this purpose, it is useful to confine oneself to the current (synchronic), graphemic
representation of words, especially because graphemic and phonemic structures of words are
not very different in Swedish. A learner of Swedish cannot be expected to deal with the
details of Swedish phonology and language history before he can finally start learning
Swedish. Hellberg 1978 takes the same view in the field of language parsing research.
Therefore, my investigation is not based on generative phonology, as in Kiefer 1970, Kiefer
1975 and Linell 1972, although I get some advantage from their results (see Section 3.4). I
pursue the classical aims of language description, standardization and instruction, as in
Collinder 1974.
In this tradition, my paper tries to make the situation easier for the learner of the Swedish
verbal system. I cannot eliminate its difficulties, of course. As pointed out in the introduction,
the help intended cannot be given with the usual catalogs of morphological irregularities in
grammar books. It is inevitable to go beyond them and to thoroughly examine the structure of
verb groups (with the same infinitive endings) which are inhomogeneous, i. e. which
contain verbs of different conjugation classes. In the context of this paper, the word ending is
not used in the traditional linguistic dichotomy, “stem vs. ending”, but in a technical sense,
meaning the last letters of a word. The number of letters in an ending is defined by pragmatic
reasons and varies between different verb groups. As a result of this investigation, I can show
the language learner in which cases analogy rules are correct and in which cases they are
wrong. As a consequence, he will know in which cases he can assign verbs to
graphematically similar pattern verbs.
In addition, the learner of the Swedish verbal system has to remember the conjugation
classes of all these pattern verbs. The latter knowledge is not represented by conjugation class
numbers, but by a couple of key forms. From Latin grammatical description, it is obvious
that a limited quantity of four key forms is sufficient to derive all the other forms of a verb (e.
g., invado, invasi, invasum, invadere). The mathematical theory behind this grammatical
principle is developed in Holl 1988. The same principle is used in grammatical descriptions
of Germanic languages (e. g., in English: go, went, gone; in German: gehen, ging, gegangen).
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
3
SAOL 1998 shows that this principle applies to Swedish as well (e. g., gå, gick, gått). Present
infinitive (I briefly define infinitive in this paper), past tense and supine are used as key
forms. In exceptional cases, these are extended by present tense and perfect participle.
All this knowledge is necessary for the language learner:
1 the applicability of analogy rules and the assignment of verbs to pattern verbs and
2 the key forms of the pattern verbs.
This complete knowledge is presented for the first time for the entire Swedish verbal system
in my verb tables (Chapter 7).
The ideas for my approach were first published in Holl 1988. There, they were checked for
Latin and for six Romance languages. Furthermore, the individual models for different
languages can easily be corrected in case of errors and adapted to language changes. This is
confirmed in the book review Schweiger 1990: 240.
3 Motivations for a reverse order presentation
3.1 The intuitive analogy rules mentioned in the introduction are induced by similarities
between infinitives when compared in reverse order.
3.2 All the compound verbs of a simple verb are automatically listed at the same place in a
verb catalog in reverse order. This is useful as the compound verbs mostly follow the same
conjugation as the corresponding simple verb.
3.3 In Holl 1988, the effectiveness of a reverse order presentation is shown for the verb
systems of Latin and six Romance languages. An essential reason is that (Neo-)Latin
infinitive endings are responsible for the assignment to a conjugation class. Thus, it is
possible to considerably reduce the linguistic material recorded in traditional verb tables.
3.4 The problem of the distribution of weak verbs to the 1st and 2nd conjugations was first
systematically discussed in Linell 1972. He mentions an interesting phenomenon: verbs with
polysyllabic roots and verbs with special final consonant clusters before the infinitive a
always belong to the 1st conjugation and never to the 2nd (Linell 1972: 67-69). The latter is
partly due to accident, partly to phonotactical reasons: a verb such as samla cannot form a 2nd
conjugation past tense *samlde, as the produced consonant cluster mld is inadmissible in
Swedish. These verbs cannot belong to the 4th conjugation either, as 4th conjugation verbs
mostly have single or geminate final consonants in the infinitive. Linell decribes this
phenomenon with his consonant cluster rule. Linells results are more or less copied by
Kiefer 1975: 139-142. The phonological details are not relevant in the context of my
problem. The only important fact is that there are homogeneous 1st conjugation verb groups
with the same final consonant cluster. Therefore, I know where I need not look for
inhomogeneous verb groups and the language learner gets some help by his intuitive sense
for pronouncability. Linells rule also induces an investigation of Swedish verbs in reverse
order, as you can find all verbs with a given final consonant cluster at the same place in a
verb catalog in reverse order.
4 How to use the verb tables
The application of my verb tables is discussed before their composition (see Chapter 5).
This is because it is easier to understand their derivation from the linguistic facts if you are
familiar with their use.
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
4
4.1 Algorithm for the assignment of an arbitrary verb to its pattern verb
My verb tables only contain pattern verbs. Their trailing parts which are not underscored
serve as analogy bases for other verbs. The analogy base plus its underscored leading part is
the pattern verb searched for. Trailing passive voice -s is ignored when applying the
following algorithm.
When you search the analogy base for an arbitrary verb, you use a simple, right-bound,
longest matching algorithm on the infinitives. It is a simple algorithm which you can use
without having any idea of computer science. You just have to find the longest possible
analogy base in the verb tables. It has two significant qualities: it has all its letters in common
with the trailing letters of the arbitrary verb and there are no longer analogy bases in the verb
tables. The first quality can also be expressed in other words: the arbitrary verb has to contain
at least as many letters as its analogy base, it has to be longer (see Section 5.3.3 for an exact
definition) than its analogy base.
In detail, the algorithm runs as follows: you start looking to see whether the entire arbitrary
verb is in the verb tables (in reverse order), either as an analogy base (not underscored), or as
a verb which is completely underscored (see exception below). If you could not find the
arbitrary verb this way, you drop its first letter. Then you look to see whether the rest is an
analogy base in the verb tables. Otherwise, you drop its second letter, and so on, until you
find the longest possible analogy base for the arbitrary verb.
The algorithm is now illustrated with a few examples:
The analogy base of skrida is rida, as skrida is longer than rida, and as any longer analogy
base matching skrida cannot be found in the verb tables. As there is no underscored part in
rida, analogy base and pattern verb are equal. The same rule applies for sprängrida, strida,
vrida, förvrida, etc.
The analogy base for sjuda is juda (this is not a Swedish verb!), as sjuda is longer than juda
and as any longer analogy base matching sjuda cannot be found in the verb tables. juda is the
part of bjuda which is not underscored, so bjuda is the pattern verb for sjuda.
Using an analogous argumentation, the analogy base of dagas is a (this is not a Swedish
verb!) as the trailing passive voice s (even ss such as in slåss, lyss) is ignored. a is contained
in jobba, so jobba is the pattern verb for dagas.
Exception: Pattern verbs which are completely underscored cannot be used as analogy
bases (or pattern verbs) for any other verb; they are valid only for themselves.
Example: indra is not the pattern verb for any other verb, in particular, not for hindra, lindra,
glindra, tindra, which are all assigned to jobba.
Remark: The arbitrary verb in question must contain at least as many letters as the analogy
base in the verb tables, e. g., förlisa is not an analogy base för lisa, and krympa not for ympa.
4.2 Algorithm for the derivation of other forms of an arbitrary verb
The principle of key forms was described in Chapter 2. I use mere analogy rules for
deriving the key forms of arbitrary verbs from the key forms of pattern verbs.
Example: rida is the pattern verb for skrida and has the key forms rida, red, ridit. Therefore,
skrida has the key forms skrida, skred, skridit. See Section 5.2 for details.
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
5
5 The derivation of the verb tables
5.1 The sources for the linguistic material
The 1st and the 3rd conjugations can be considered as regular: the 1st for the verbs ending in
-a, the 3rd for those ending in another vowel. Thus, when compiling the linguistic material,
you must look for catalogs comprising the rest, i. e. the 2nd and 4th conjugations, in order to
figure out inhomogeneous verb groups. It is quite easy to find lists with all the irregular verbs
(complete 4th conjugation and parts of the 2nd conjugation). They are included in every
grammar book and every dictionary. It is very difficult, however, to find complete catalogs
with all regular verbs in the 2nd conjugation. Even SAG 1999 does not contain any, but only
single examples and lists with exceptions.
The first complete list of all verbs in the 2nd and 4th conjugations was published in
Collinder 1974: 83-99 under the headline “Tempusböjningslista”. However, Swedish has
changed since 1974. Thus, this catalog has become a bit obsolete and does no longer
represent the current linguistic state described in SAOL 1998.
The second useful source for my research is Hellberg 1978. Hellberg presents lists of
equally inflected verbs. “The dictionary is by no means exhaustive, but large enough ... to
give an idea of the distribution of paradigms in a basic Swedish vocabulary.” (Hellberg 1978:
12).
The third source, Odhner 1979, is very important. This reverse dictionary classifies all
words according to parts of speech and according to inflectional classes. Thus, you find
information on inhomogeneous verb groups with the same infinitive ending.
A simple attempt with a similar aim as mine was done in Perridon 1985. It contains a
couple of interesting ideas, but it does not exceed the state of a mere attempt, as essential
principles of software engineering are violated. Perridon’s algorithm for the production of
verb forms is written as a Pascal program without any published design concept and without
any comments in the source code. It is very complex and only documented in some fragments
so that it cannot be followed. The worst critique is that it contains language data in the form
of program constants so that the correction of errors and the adaptation to language changes
would require modifications of the algorithm - a terrible job. Furthermore, one assertion
shows that the linguistic facts were not analyzed correctly: Perridon 1985: 95 states that verbs
ending in -örja follow the 2nd conjugation. This is wrong as the frequent verb börja belongs
to the 1st conjugation. For the reasons mentioned, the results presented are useless for
language instruction and I did not make any further use of this paper.
In contrast to Perridon 1985, my approach is completely documented and efficient. My
algorithm does not depend on the language considered. Underlying language data are used in
variables, whose values are separated (see Section 5.3, remark 1). Therefore, necessary
changes do not affect my algorithm, but only the language data which are accessible to easy
modification.
The material from the sources mentioned was checked with SAOL 1998. Every verb group
which might contain non-1st conjugation verbs (according to Linells consonant cluster rule)
was examined in detail with the reverse dictionary Allén 1993 and SAOL 1998.
5.2 The verbs and verb forms mentioned
I exclude verbs marked as ‘finlandsvensk’ (Finland Swedish), ‘sydsvensk’ (Southern
Swedish) or ‘provinsiellt’ (local, regional) in SAOL 1998.
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
6
Compound verbs are normally included in my investigation if they are recorded both in
SAOL 1998 and in Allén 1993. Only if the conjugation of a simple verb is determined by
differences in meaning (such as sluta, see introduction), its compounds are excluded. This is
because their conjugation is determined by the same difference in meaning.
Passive and deponent forms ending in -s are treated in the reverse order, as if the trailing s
would not exist.
Furthermore, SAOL 1998 classifies some alternative verbs and verb forms with 'även'
(rekommendation i andra hand), i. e. they are only recommended in second hand. Other verbs
and verb forms are marked as ‘ålderdomligt’ (obsolete). Both are either not considered or
recorded in parentheses in my verb tables. If a verb form is mentioned in SAG 1999 and not
in SAOL 1998, I confine myself to the more restrictive standard of SAOL 1998.
My verb tables in Chapter 7 contain four columns:
Column 1: present infinitive
Column 2: present tense, if necessary, i. e. if it does not follow the rule:
If present infinitive in -a and supine in -at, then present tense in -ar.
If present infinitive in -a and supine not in -at, then present tense in -er.
If present infinitive in -V (vowel not = a), then present tense in -Vr.
Column 3: past tense
Column 4: supine
Column 5: the verb’s meaning in a Swedish paraphrase. It is mentioned if and only if it is
relevant for the verb’s conjugation. This applies only for verbs with different conjugations
depending on their meaning, such as sluta which was mentioned in the introduction.
As discussed in Chapter 2, the key forms in columns 1 to 4 are sufficient to derive every
other verb form. The derivation procedures can be read in every Swedish grammar book.
Therefore, I do not deal with imperative, present participle, past participle, compound tenses
and passive voice forms. Subjunctive and optative forms are excluded from my investigation
as well because they have become rare in everyday language.
5.3 Systematization of the linguistic material
The systematization is done in the following three steps. Due to the huge amount of
linguistic material, I cannot show it in detail for each verb group. I only illustrate it with one
example.
5.3.1 All the 2nd and 4th conjugation verbs are sorted in reverse order. Thus, verb groups
with the same graphematic ending arise automatically.
Example: bända, hända, lända, sända, tända, vända constitute the group -ända.
5.3.2 The verb groups are completed with 1st and 3rd conjugation verbs. This is done with
reverse dictionaries, such as Allén 1993 and Odhner 1979.
Example: ända, skända, blända, fullända.
5.3.3 The most simple and efficient analogy rule for each group is stated.
Example: “In general, the group -ända follows the 1st conjugation, but bända, hända, lända,
sända, tända, vända and all longer verbs follow the 2nd conjugation.”
In this context, I use the following convention: a verb A is ‘longer’ than another verb B in
reverse order if verb B is contained at the end of verb A. Example: anlända, blända, fullända
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
7
are longer than lända. Equal length is also included in the mathematical sense, so lända is
longer than lända as well.
That’s why the above rule contains a mistake in its preliminary form: it would induce the
wrong proposition that blända and fullända would follow the same conjugation as lända.
Thus, the above rule has to be completed by: “blända and fullända belong to the 1st
conjugation.” By the way, ända and skända are assigned to the pattern verb jobba of the 1st
conjugation (cf. Section 4.1).
In formal terms, the rule is written like this in my verb tables:
bända
bände
bänt
hända
hände
hänt
lända
lände
länt
blända
bländade
bländat
fullända
fulländade
fulländat
sända
sände
sänt
tända
tände
tänt
vända
vände
vänt
Remark 1: This rule (and the entire verb tables) is not part of the algorithm which assigns
an arbitrary verb to its pattern verb (see Section 4.1). It is part of the language-specific data
the algorithm uses.
Remark 2: At first glance, it may seem strange that I do not treat fullända as a (linguistic)
compound of ända, but just as a verb (technically) longer than lända. Otherwise, I would
expect the user of my verb tables to have the ability to decompose all the Swedish compound
verbs. This would require a lot of knowledge about Swedish morphological structures: a list
of possible verbal prefixes and of possible verbal stems. However, this is native speaker
knowledge, which a language learner does not have. I will just give two crucial examples: Is
presslägga a compound to lägga or to slägga? The first alternative is correct. Does påta
belong to ta (4th conjugation)? Yes, but there is also a simple verb påta, ‘to dig’ (1st
conjugation).
This leads to my first principle: treat all verbs (in particular their infinitives) without
respect to their internal morphological structure; just consider them as unstructured strings of
letters. Thus, the desriptions of the verbal systems become a lot easier. This point of view
was already used successfully in Holl 1988.
Remark 3: Theoretically, the exception to the above rule could also have been formulated
this way: “Verbs longer than lända do not follow the 2nd conjugation, with the exception of
its compounds (e. g. anlända).” This would have the following disadvantage:
I would be obliged to present complete lists of compound verbs, which is a very difficult
task. Even SAOL 1998 does not record all compound verbs. Therefore, I prefer to record the
longer 1st conjugation verbs, in the example: blända and fullända.
This leads to my second principle: avoid lists of compound verbs as far as possible. I
cannot obey this principle only in the case of the compounds of short non-1st conjugation
verbs, such as ta, äta, dra. If I chose ta as a pattern verb for verbs longer than ta, I would
indeed avoid listing its compounds, but I would have to list all 1st conjugation verbs ending in
-ta. Therefore, it is better to obey my lower third principle in this case: I exclude ta as pattern
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
8
verb, have to list its compounds, but avoid listing the huge quantity of all 1st conjugation
verbs ending in -ta.
Remark 4: Theoretically, the above rule could also have been formulated this way: “In
general, the group -ända follows the 2nd conjugation, but ända, skända, blända follow the 1st
conjugation.” There were two disadvantages, however:
1. There is a general rule in Swedish that verbs mostly follow the 1st conjugation, as it
comprises the majority of verbs. With this theoretical rule, however, the language learner
would learn an explicit list of 1st conjugation verbs which follow the general rule, but he
would only get an implicit knowledge of the exeptions of the general rule. Figures from SAG
1999: vol. 2, pg. 558, however, show how important non-1st conjugation verbs are in modern
Swedish (and therefore explicit knowledge about them): in newspaper texts from the 1960s,
only 25 % of the occurring verb forms belong to 1st conjugation verbs, which include 67 % of
all different Swedish verbs. In terms of computer linguistics: 1st conjugation verbs amount to
67 % of the verbal types, but only 25 % of the verbal tokens.
2. My verb tables shall be open for changes in the Swedish language, but the necessity of
modifying them should not arise very often. The 1st conjugation is the only productive one in
Swedish, that is, the only one to which new verbs are assigned. If I were to record many
enumerations of 1st conjugation verbs, I would frequently run the risk to have to include new
ones when they arise in Swedish. If I try to minimize those enumerations, I only run the risk
of new verbs which are longer than 2nd or 4th conjugation verbs. Example: I would have to
include a fictive 1st conjugation verb *flägga which would otherwise be assigned to the 4th
conjugation verb lägga.
This leads to my third principle (subordinate to my second principle): avoid lists of 1st
conjugation verbs as far as possible. I cannot obey this principle in two cases: firstly, if the
quantity of 1st conjugation verbs is very small in comparison with the other verbs of a verb
group; e. g., all verbs ending in -öja belong to the 2nd conjugation with the only exception
slöja. Secondly, if I would violate my higher second principle; e. g., I do not list the
compounds of the 2nd conjugation verbs lända, tiga, åka etc., but present the very short lists
of longer 1st conjugation verbs instead.
6 Additional results
There are a couple of additional results which I obtain from compiling my verb tables.
They concern types of (in)homogeneous groups (6.1) and a short comparison with Latin and
Romance linguistic facts (6.2).
6.1 Homogeneity of verb groups with the same infinitive ending
A Swedish verb group with the same infinitive ending can be homogeneous (its verbs
belong to the same conjugation class) or inhomogeneous (its verbs belong to different ones).
It is more likely to be homogeneous, the more trailing letters define it. Thus, the group ending
in -binda is homogeneous, but it contains only the simple verb binda and its compounds. As
such a result is not interesting, I do not mention verb groups of that kind.
6.1.1 Homogeneous groups
1st conjugation: There are a lot of verb groups that contain only verbs of the 1st conjugation,
according to Linells consonant cluster rule. A complete list can be found in Linell 1972: 6769 (cf. Section 3.4).
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
9
2nd conjugation: There are only a few homogeneous groups with verbs of the 2nd
conjugation: e. g., all verbs ending in -länga, -ränka, -ärka, -räka, -änna, -ärpa and a few
which contain only two verbs.
3rd conjugation: All verbs ending in -o, -y, and -ä follow the 3rd conjugation.
4th conjugation: There are only a few homogeneous groups: all verbs ending in -juta, -ryta.
6.1.2 Inhomogeneous groups
1st and 2nd conjugation: e. g. verbs ending in -eda etc.
1st and 4th conjugation: e. g. verbs ending in -inda etc.
1st, 2nd and 4th conjugation: e. g. verbs ending in -ida, -ippa etc.
2nd and 4th conjugation: e. g. verbs ending in -lippa etc.
3rd and 4th conjugation: e. g. verbs ending in -e, -å and -ö.
The rest of the possible combinations does not occur, as verbs ending in -a (candidates for
the 1st and 2nd conjugation) can not occur together with verbs ending in another vowel
(candidates for the 3rd conjugation) in the same group.
6.2 Comparison with (Neo-)Latin languages
As the Swedish verbal system contains a lot of inhomogeneous verb groups, in contrast to
(Neo-)Latin languages, the condensation of the linguistic material in Section 5.3.3 is less
effective in Swedish. My Swedish verb tables contain about 550 pattern verbs with 3 key
forms each (with no respect to present tense). This amounts to 1650 entries.
According to Holl 1988: 181 only Romanian needs more pattern verbs than Swedish,
namely 800. French and Portuguese need only 150. French requires 7 key forms and
Portuguese 6, twice as many as Swedish.
A final interesting detail: the Latin verb tables in Holl 1988: 204-215 contain about 400
pattern verbs with 4 key forms each. This amounts to 1600 entries. Thus, I can state: with
regard to the assignment of verbs to pattern verbs, Swedish is more difficult than Latin.
7 Verb tables with pattern verbs and key forms
The following conventions are used for character attributes:
all pattern verbs and key forms of the 1st and 3rd conjugations: normal
regular key forms of the 2nd conjugation: bold type;
irregular key forms of the 2nd conjugation: italics in bold type;
key forms of the 4th conjugation: underscored bold type.
letters;
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
jobba
jobbade
jobbat
leda
leda
ledas
reda
reda
breda
freda
sveda
ledade
ledde
leddes
redade
redde
bredde
fredade
svedde
ledat
lett
letts
redat
rett
brett
fredat
svett
idas
lida
smida
gnida
iddes
led
smidde
gned
itts
lidit
smitt
gnidit
rida
sprida
bestrida
red
(spridde) spred
bestred
kved
svidade
sved
kvidit
svidat
svidit
binda
band
bundit
varda
(tordas)
bände
hände
lände
bländade
fulländade
sände
tände
vände
varder
--
röra sig i en led, böja
föra, vara främst
känna leda
idka rederigörelse
göra i ordning
ridit
spritt, spridit
bestritt, bestridit
kvida
svida
svida
bända
hända
lända
blända
fullända
sända
tända
vända
10
vart
--
klä
göra ont
bänt
hänt
länt
bländat
fulländat
sänt
tänt
vänt
(ptcp. vorden) -(tordats) i första hand: töras
bjuda
ljuda
ljuda
bjöd
ljudade
ljöd
bjudit
ljudat
ljudit
lyda
tyda
löd, lydde
tydde
lytt
tytt
låda
råda
lådde
rådde
lått
rått
(kläda)
späda
rädas
klädde
spädde
räddes
klätt
spätt
rätts
uttala ljud för ljud
ge ljud ifrån sig, höras
i första hand: klä
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
skräda
träda
kväda
skrädde
trädde
(kvädde) kvad
skrätt
trätt
kvädit
öda
föda
göda
löda
flöda
föröda
(stöda)
ödde
födde
gödde
lödde
flödade
förödde
stödde
ött
fött
gött
lött
flödat
förött
stött
gnaga
(draga)
(taga)
staga
förstaga
gnagde
drog
tog
stagade
förstagade
gnagt
dragit
tagit
stagat
förstagat
låg
pliggade
tiggde
legat
pliggat
tiggt
hugga
högg
huggit
bygga
brygga
brygga
byggde
bryggade
bryggde
byggt
bryggat
bryggt
lägga
slägga
(la) lade
släggade
lagt
släggat
niga
tiga
beriktiga
berättiga
viga
föreviga
neg
teg
beriktigade
berättigade
vigde
förevigade
nigit
tigit
beriktigat
berättigat
vigt
förevigat
ringa
ringade
ringa
ringde
bringa bringar bringade,bragte
springa
sprang
förringa
förringade
stinga
-tvinga tvingar (tvang) tvingade
ringat
ringt
bringat, bragt
sprungit
förringat
stungit
(tvungit) tvingat
ligga
pligga
tigga
sjunga
ligger
sjöng
sjungit
tynga
tyngde
tyngt
dänga
dängde
dängt
i första hand: stödja
i första hand: dra
i första hand: ta
brygga över
brygga kaffe
förse med ring
ljuda, telefonera
11
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
hänga
länga
mänga
tränga
stränga
omstränga
stänga
svänga
hängde
längde
mängde
trängde
strängade
omsträngade
stängde
svängde
duga
ljuga
suga
(dugde) dög
ljög
sög
blygas
flyga
smyga
blygdes
flög
smög
äga
säga
väga
ha
coacha
smasha
(bedja)
har
(beder)
hängt
längt
mängt
trängt
strängat
omsträngat
stängt
svängt
dugt
ljugit
sugit
blygts
flugit
smugit
ägde
(sa) sade
vägde
ägt
sagt
vägt
hade
coachade
smashade
haft
coachat
smashat
bad
bett
gladde
(stadde)
glatt
(statt)
stödde
stött
leja
lejde
lejt
skilja
vilja
(skiljde) skilde
vill
ville
(skiljt) skilt
velat
glädja
gläder
(städja)(städ(j)er)
stödja
sälja
tälja
välja
dväljas
kvälja
kvälja
dölja
följa
hölja
skölja
tämja
vämjas
stöder
sålde
täljde
valde
dvaldes, dväljdes
kväljde
kvalde
dolde
följde
höljde
sköljde
tämjde
vämjdes
sålt
täljt
valt
dvalts, dväljts
kväljt
kvalt
dolt
följt
höljt
sköljt
tämjt
vämjts
i första hand: be
inge äckel
obehörigt klandra
12
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
tänja
vänja
tänjde
vande
tänjt
vant
skönja
skönjde
skönjt
snärja
värja
(svärja)
svär
besvärja -svärjer
smörja
spörja
sörja
snärjde
värjde
svor
besvor
(besvärjde)
snärjt
värjt
svurit
besvurit
(besvärjt)
(smörjde) smorde
sporde
sörjde
(smörjt) smort
sport
sörjt
sade
väjde
sagt
väjt
böja
slöja
böjde
slöjade
böjt
slöjat
klicka
klicka
dricka
spricka
sticka
sticka
klickade
klack
drack
sprack
stickade
stack
klickat
klickat
druckit
spruckit
stickat
stuckit
lyckas
lycka
knycka
rycka
tycka
stycka
misstycka
lyckades
lyckte
knyckte
ryckte
tyckte
styckade
misstyckte
lyckats
lyckt
knyckt
ryckt
tyckt
styckat
misstyckt
läcka
kläcka
släcka
smäcka
knäcka
räcka
träcka
träcka
täcka
väcka
läckte
kläckte
släckte
smäckte
knäckte
räckte
träckade
träckte
täckte
väckte
läckt
kläckt
släckt
smäckt
knäckt
räckt
träckat
träckt
täckt
väckt
leka
smeka
steka
lekte
smekte
stekte
lekt
smekt
stekt
(säja)
väja
13
i första hand: svära
i första hand: säga
lägga en klick; knäppa
spritta av sinnesrörelse
sticka strumpor
ge ett stick
ha framgång
stänga
avge träck
dra fartyg
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
beveka
bevekte
bevekt
förlika
snika
skrika
vika
svika
förlikte
snikte, snek
skrek
vek
svek
förlikt
snikt, snikit
skrikit
vikt, vikit
svikit
slinka
stinka
slank
stank
slunkit
--
sjunka
sjönk
sjunkit
skänka
blänka
dränka
sänka
tänka
skänkte
blänkte
dränkte
sänkte
tänkte
skänkt
blänkt
dränkt
sänkt
tänkt
kokade
kokte, kokade
kokat
kokt, kokat
styrka
styrkte
styrkt
märka
märkte
märkt
sluka
(slök) slukade
slukat
byka
dyka
ryka
ryka
stryka
bykte
dök
(rykte) rök
rök
strök
bykt
dykt
rykt
rykt
strukit
åka
pjåka
råka
åkte
pjåkade
råkade
åkt
pjåkat
råkat
läka
späka
bräka
kväka
läkte
späkte
bräkte
kväkte
läkt
späkt
bräkt
kväkt
röka
föröka
söka
rökte
förökade
sökte
rökt
förökat
sökt
gala
gal
gol
mala (maler) mal
malde
tala
talar (-lte) talade
betala
-talar (-lte) talade
galt, galit
malt
(talt) talat
betalt, betalat
koka
koka
kokar
14
klumpa (jord)
bringa i kokning
sända ut rök
börja slåss; gå förlorat
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
falla
befalla
föll
befallde
fallit
befallt
spilla
spillde
spillt
fylla
förgylla
skylla
fyllde
förgyllde
skyllde
fyllt
förgyllt
skyllt
hålla
hushålla
höll
hushållade
hållit
hushållat
fälla
gälla
hälla
skälla
smälla
smälla
gnälla
drälla
ställa
välla
kvällas
kvälla
tillvälla
skola
skola
fällde
gällde
hällde
skällde
smällde
small, smällde
gnällde
drällde
ställde
vällde
kvällades
kvällde
tillvällade
ska(ll)
kyla
skyla
skyla
fällt
gällt
hällt
skällt
smällt
smällt
gnällt
drällt
ställt
vällt
kvällats
kvällt
tillvällat
skolat
skolat
utbilda, kila med skol
komma att, böra
kylde
skylade
skylde
kylt
skylat
skylt
sätta i skyl
hölja, dölja
tål
tålde
tålt
stjäla
anmäla
genmäla
stjäl
stal
anmälde
(-mälte) -mälde
stulit
anmält
genmält
förnam
(sam) simmade
kom
förnummit
(summit) simmat
kommit
rymma
förgrymmas
dämma
skämma
klämma
drämma
stämma
simmar
bli kväll
välla, flöda
skolade
skulle
tåla
förnimma
simma
komma
slå med en smäll
ge knallande ljud
rymde
förgrymmades
dämde
skämde
klämde
drämde
stämde
rymt
förgrymmats
dämt
skämt
klämt
drämt
stämt
15
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
gömma
glömma
drömma
berömma
tömma
gömde
glömde
drömde
berömde
tömde
gömt
glömt
drömt
berömt
tömt
värma
svärma
värmde
svärmade
värmt
svärmat
gräma
grämde
grämt
döma
dömde
dömt
(mente) menade
förmenade
-menar (-mente) -menade
(ment) menat
förmenat
(-ment) -menat
mena
förmena
förmena
menar
skina
vina
sken
ven
nämna
nämnde
finna
hinna
minnas
spinna
rinna
skrinna
vinna
tvinna
utvinna
kunna
fann
hann
mindes
spann
rann
skrinnade
vann
tvinnade
utvann
kan
nämnt
funnit
hunnit
mints
spunnit
runnit
skrinnat
vunnit
tvinnat
utvunnit
kunnat
begynna
begynte
begynt
känna
kände
känt
bryna
syna
synas
brynte
synade
syntes
brynt
synat
synts
röna
dröna
skapa
svepa
knipa
pipa
skapar
neka, förvägra
anse
skinit
vinit
kunde
syns
16
rönte
drönade
rönt
drönat
skapade
svepte
skapt, skapat
svept
knep
pipade
knipit
pipat
besiktiga
vara synlig, ses, tyckas
rörförmigt vecka
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
pipa
gripa
pep
grep
pipit
gripit
hjälpa
hjälpte
hjälpt
dimpa
damp
dumpit
krympa
krympte
krympt
klippa
slippa
klippte
slapp
klippt
sluppit
släppa
knäppa
skräppa
täppa
släppte
knäppte
skräppte
täppte
släppt
knäppt
skräppt
täppt
skärpa
skärpte
skärpt
snörpa
snörpte
snörpt
supa
stupa
söp
(stöp) stupade
supit
stupat
nypa
drypa
strypa
nöp
dröp
(strypte) ströp
dräpa
dräpte
köpa
förlöpa
gröpa
gröpa
fara
befara
befara
spara
vara
vara
förevara
övervara
närvara
svara
dra
neddra
bedra
föredra
segdra
(nupit) nypt
drupit
strypt
dräpt
köpte
köpt
förlöpte (förlupit) förlöpt
gröpade
gröpat
gröpte
gröpt
far
befar
spar(ar)
är
-varar
-varar
drar
-drar
-drar
-drar
-drar
for
befarade
befor
sparade
varade
var
förevar
övervarade
närvarade
svarade
drog
neddrog
bedrog
föredrog
segdrog
kvittra, gnälla, vina
farit
befarat
befarit
sparat
varat
varit
förevarit
övervarat,-varit
närvarat, -varit
svarat
dragit
neddragit
bedragit
föredragit
segdragit
grovmala
gräva, urholka
frukta
fara på/över
pågå; utsöndra var
hjälpverb
17
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
bidra
tilldra
framdra
andra
undandra
sammandra
indra
fråndra
uppdra
överdra
hårdra
fördra
bortdra
avdra
ådra
ådra
yra
yra
hyra
pyra
styra
bära
begära
(skära)
skära
beskära
beskära
oskära
lära
nära
tära
svära
besvära
böra
föra
göra
höra
köra
köra
sköra
snöra
röra
*töra
töras
störa
störa
-drar
-drar
-drar
-drar
-drar
-drar
-drar
-drar
-drar
-drar
-drar
-drar
-drar
-drar
-drar
yr
hyr
pyr
styr
bär
begär
skär
-skär
-skär
lär
när
tär
svär
bör
för
gör
hör
kör
snör
(snörar)
rör
tör
törs
stör
bidrog
tilldrog
framdrog
androg
undandrog
sammandrog
indrog
fråndrog
uppdrog
överdrog
hårdrog
fördrog
bortdrog
avdrog
ådrade
ådrog
bidragit
tilldragit
framdragit
andragit
undandragit
sammandragit
indragit
fråndragit
uppdragit
överdragit
hårdragit
fördragit
bortdragit
avdragit
ådrat
ådragit
yrade
yrde
hyrde
pyrde
styrde
yrat
yrt
hyrt
pyrt
styrt
bar
begärde
(skärade)
skar
beskärde
beskar
oskärade
lärde
närde
tärde
svor
besvärade
burit
begärt
(skärat)
skurit
beskärt
beskurit
oskärat
lärt
närt
tärt
svurit
besvärat
borde
förde
gjorde
hörde
körade
körde
skörade
snörde
(snörade)
rörde
torde
tordes
störade
störde
bort
fört
gjort
hört
körat
kört
skörat
snört
(snörat)
rört
tort
(tordats) torts
störat
stört
18
göra ådrig
få en sjukdom
tala förvirrat
virvla,drivas med vinden
rena
klippa
skänka
klippa (träd)
sjunga i bakgrundskör
styra, åka
stödja med störar
besvära, oroa
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
resa
reste
rest
gläfsa
gläfste
gläfst
fisa
förlisa
fes
förliste
fisit
förlist
frälsa
frälste
frälst
glänsa
glänste
glänst
näpsa
näpste
näpst
kyssa
kysste
kysst
förtjusa
förtjuste
förtjust
lysa
plysa
mysa
nysa
fnysa
pysa
rysa
frysa
frysa
lyste
plysade
(mös) myste
(nyste) nös
fnös, fnyste
(pös) pyste
rös, ryste
frös
frös, fryste
lyst
plysat
myst
nyst
fnyst
pyst
ryst
frusit
frusit, fryst
låsa
flåsa
låste
flåsade
låst
flåsat
jäsa
fjäsa
läsa
snäsa
fräsa
väsa
jäste
fjäsade
läste
snäste
fräste
väste
jäst
fjäsat
läst
snäst
fräst
väst
ösa
slösa
rösa
överösa
öste
slösade
rösade
överöste
öst
slösat
rösat
överöst
ta
tar
tog
tagit
-tar
-tar
-tog
-tog
-tagit
-tagit
medta
medtar
medtog
medtagit
beta
beta
betar
tillfångata
tillvarata
betade
betog
betat
betagit
19
stelna av köld [intrans]
bevara gm köld [trans.]
äta gräs; bryta; betsa
beröva, överväldiga
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
arbeta
företa
heta
veta
arbetade
företog
hette
visste
arbetat
företagit
hetat
vetat
gifta
gifte
gift
lyfta
klyfta
lyfte
klyftade
lyft
klyftat
beslagta
företar
heter
vet
-tar
beslagtog
bita
bita
skita
skita
slita
smita
delta
tillta
bitade
bet
skitade
sket
slet
smet
deltar
tilltar
smälta
smälta
välta
välta
svälta
svälta
deltog
tilltog
smälte
(smalt) smälte
vältade
välte
svalt
(svalt) svälte
beslagtagit
bitat
bitit
skitat
skitit
slitit
smitit
smält
(smultit) smält
vältat
vält
svultit
svält
framtar
framtog
framtagit
anta
undanta
(genta)
inta
slinta
frånta
antar
-antar
(gentar)
intar
fråntar
antog
undantog
(gentog)
intog
slant
fråntog
antagit
undantagit
(gentagit)
intagit
sluntit
fråntagit
uppta
upptar
upptog
upptagit
överta
förta
övertar
förtar
övertog
förtog
övertagit
förtagit
mista
rista
rista
brista
drista
turista
mister
(-ade) miste
ristade
(-ade) riste
brast
dristade
turistade
missta
misstar
misstog
dela i bitar
nafsa; vara skarp
smutsa ner
tömma tarmen
deltagit
tilltagit
framta
rister
20
(mistat) mist
ristat
(ristat) rist
brustit
dristat
turistat
misstagit
göra flytande [trans.]
bli flytande [intrans.]
bearbeta med vält
falla åt sidan, stjälpa
hungra [intrans.]
låta hungra [trans.]
skära, hugga
skära (smärta); skaka
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
*måsta
måste
måste
måst
fästa
befästa
gästa
nästa
fäster
(-ade) fäste
befäste
gästade
näste, nästade
(fästat) fäst
befäst
gästat
näst, nästat
vetta
vetter
vette
vettat
gitta
spritta
sitta
gitter
gitte
spratt
satt
gittat
-suttit
iaktta
motta
bortta
utta
iakttar
mottar
borttar
uttar
iakttog
mottog
borttog
uttog
iakttagit
mottagit
borttagit
uttagit
hytta
hytte
hytt
sprätta
sprätta
sätta
skvätta
skvätta
sprättade
sprätte
satte
skvätte
(skvatt) skvätte
sprättat
sprätt
satt
skvätt
skvätt
gjuta
göt
gjutit
sluta
slutar
slutade
slutat
sluta
slutar slöt, slutade
slutat
slutit
sluta
sluter
slöt
besluta
-slutar -slöt,-slutade (-slutit) beslutat
avsluta
-slutar
avslutade
avslutat
(avsluta) (-sluter)
(avslöt)
(avslutit)
avta
avtar
byta
flyta
åta
låta
plåta
påta
påta
gråta
äta
(förgäta)
fläta
släta
mäta
åtar
påtar
avtog
ta upp; skära upp
krafsa; vara sprättig
stänka [personligt]
stänka [opersonligt]
komma till ett slut
resultera i
stänga; dra slutsats
slutföra; upphöra
överenskomma om
avtagit
bytte
flöt
bytt
flutit
åtog
lät
plåtade
påtade
påtog
grät
åtagit
låtit
plåtat
påtat
påtagit
gråtit
åt
-flätade
slätade
mätte
ätit
(förgätit)
flätat
slätat
mätt
gräva; peta; pyssla
överta
21
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
näta
räta
fräta
träta
uträta
täta
tröstäta
väta
aväta
nätade
rätade
frätte
trätte
uträtade
tätade
tröståt
vätte
avåt
nätat
rätat
frätt
trätt
uträtat
tätat
tröstätit
vätt
avätit
höta
sköta
blöta
möta
nöta
stöta
hötte
skötte
blötte
mötte
nötte
stötte
hött
skött
blött
mött
nött
stött
hade
skavde
(begrov) begravde
haft
skavt
begravt
(hava)
skava
begrava
leva
sleva
(giva)
(bliva)
kliva
riva
trivas
skälva
välva
sova
lever
levde
slevade
gav
blev
klev
rev
trivdes
(skalv) skälvde
välvde
sov
i första hand: ha
levat, levt
slevat
gett, givit
blivit
klivit
rivit
trivts
i första hand: ge
i första hand: bli
skälvt
välvt
sovit
ärva
nedärva
djärvas
kringvärva
omvärva
ärvde
nedärvde
djärvdes
kringvärvde
omvärvde
ärvt
nedärvt
djärvts
kringvärvt
omvärvt
yva
yvas
klyva
yvade
yvdes
klöv
yvat
yvts
kluvit
häva
gräva
kräva
väva
kväva
hävde
grävde
krävde
vävde
kvävde
hävt
grävt
krävt
vävt
kvävt
bli yvigare; yvas
vara stolt
22
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
behöva
söva
stridsöva
behövde
sövde
stridsövade
behövt
sövt
stridsövat
växa
växte
vuxit, växt
be
ge
ske
le
se
förse
bad
gav
skedde
log
såg
(-sedde) -såg
bett
gett, givit
skett
lett
sett
-sett
bli
blev
bo
fly
lyss
lyss
blivit
bodde
bott
flydde
lyddes
flytt
--
få
gå
slå
må
må
nå
stå
fick
fått
gick (ptcp. gången) gått
slog
slagit
måtte
mått
mådde
mått
nådde
nått
stått
stod
klä
klädde
klätt
dö
strö
dog
strödde
dött
strött
hjälpverb
känna sig
23
Alfred Holl: The inflectional morphology of the Swedish verb. In: Arkiv för nordisk filologi 116 (2001) 193-220
8 Bibliography
Allén, Sture & Sjögreen, Christian 1993: Norstedts svenska baklängesordbok.
Norstedts, Stockholm.
Collinder, Björn 1974: Svensk språklära. CWK Gleerup, Lund.
Hellberg, Staffan 1978: The morphology of present-day Swedish.
Word-inflection, word-formation, basic dictionary [= Data linguistica 13].
Almqvist & Wiksell, Stockholm.
Holl, Alfred 1988: Romanische Verbalmorphologie und relationentheoretische
mathematische Linguistik. Axiomatisierung und Anwendung des klassischen Wortund-Paradigma-Modells [= Linguistische Arbeiten 216]. Niemeyer, Tübingen.
Kiefer, Ferenc 1975: “Das schwedische Verbalsystem”.
Kiefer, Ferenc (ed.): Morphologie und generative Grammatik.
Athenaion, Frankfurt, pp. 129-163.
Kiefer, Ferenc 1970: Swedish morphology. Skriptor, Stockholm.
Linell, Per 1972: Remarks on Swedish morphology [= Reports from Uppsala University
Department of Linguistics Nr. 1, 1972]. Universitetet, Uppsala.
Odhner, Einar et al. 1979: Svenskt rimlexikon. Forum, Stockholm, 3rd edition.
Perridon, Harry 1985: “Verbens morfologi i skriven rikssvenska”.
Tijdschrift voor Skandinavistiek 6, pp. 88-120.
SAG 1999 = Teleman, Ulf et al.: Svenska Akademiens grammatik.
Svenska Akademien, Stockholm.
SAOL 1998 = Svenska Akademiens ordlista över svenska språket.
Norstedts, Stockholm, 12th edition.
Schweiger, Fritz 1990: “Review of Holl 1988”. Yearbook of Morphology 3, pp. 238-240.
24